- Published
- Updated
Two reports may have different study names yet include participants from the same sites and period. Partial overlap creates a dependence problem distinct from multiple reports of one study. Examine the possibility of shared data before assuming independent participants in a combined interpretation.
Scope of the decision
Sites, recruitment dates, data sources and eligibility criteria provide useful clues. A shared period alone does not prove participant overlap. If reporting is unclear, state uncertainty and seek suitable additional information where appropriate. Link the analytical approach to the likely importance of that uncertainty.
Worked review
For example, a multicentre device registry and a single-site analysis cover the same period. The local site also contributes to the registry. Rather than adding the sample sizes directly, the team investigates overlap. If details remain unavailable, it considers separate presentation or an appropriate sensitivity approach with methodological expertise.
Practical checks
- Compare sites, periods and data sources as well as study names.
- Record confirmed shared data separately from possible overlap.
- Explain how uncertain participant independence affects the analysis decision.
Acceptance and limits
Acceptance requires visible consideration of shared data and a justified analytical route. Suspected overlap does not automatically require exclusion of every relevant study. Equally, ignoring possible dependence can overstate precision, so preserve the limits of inference in the report.
Frequently asked questions
- Does a different study number establish independence?
- Not by itself; data may still originate from the same source population.
- Do shared authors prove overlap?
- They are only a clue and should be considered with participant periods and data sources.